Upload
others
View
6
Download
0
Embed Size (px)
Citation preview
2017
ANNUAL AIR
QUALITY REPORT
AIR POLLUTION CONTROL DISTRICT
SAN LUIS OBISPO COUNTY
3433 Roberto Court
San Luis Obispo, California 93401
Phone: (805) 781-5912
Fax: (805) 781-1002
Burn Advisory (toll free): (800) 834-2876
Email: [email protected]
Website: www.slocleanair.org
Air Pollution Control Officer: Gary Willey
Senior Staff
Engineering & Compliance: Dora Drexler
Planning & Outreach: Andy Mutziger
Fiscal, Administrative & Technical Services: Kevin Kaizuka
2017 Annual Air Quality Report
Published November 2018
By Fiscal, Administrative & Technical Services Division
Monitoring Section Staff:
Karl Tupper
Jaime Contreras
Table of Contents
Executive Summary ...................................................................................................................................................... 1
Air Quality Monitoring and Data ................................................................................................................................. 2
Ambient Air Pollutants Of Local Concern .................................................................................................................. 5
Ozone ...................................................................................................................................................................... 5
Particulate Matter .................................................................................................................................................. 5
Nitrogen Dioxide, Sulfur Dioxide, and Carbon Monoxide................................................................................. 6
State and National Ambient Air Quality Standards .................................................................................................. 7
Exceptional Events ................................................................................................................................................. 7
Ozone and Gaseous Pollutant Summary .................................................................................................................. 9
Maximum Values ................................................................................................................................................... 9
Visual Ozone Summary ....................................................................................................................................... 10
Particulate Matter Summary ..................................................................................................................................... 12
Highest 24-hr Concentrations and Annual Averages ...................................................................................... 12
Visual PM2.5 and PM10 Summaries ..................................................................................................................... 14
10-Year Trends ............................................................................................................................................................ 16
Ozone .................................................................................................................................................................... 16
Particulate Matter ................................................................................................................................................ 18
Ambient Air Monitoring Network Plans ................................................................................................................... 18
Appendix A: Assessing the Effectiveness of ODSVRA Mitigations ........................................................................ 23
Introduction .......................................................................................................................................................... 23
Background ........................................................................................................................................................... 23
Analysis .................................................................................................................................................................. 26
Conclusions ........................................................................................................................................................... 28
Appendix B: Ambient Respirable Crystalline Silica Monitoring ............................................................................ 29
Executive Summary & Background ................................................................................................................... 29
General Considerations ...................................................................................................................................... 29
Methodology ......................................................................................................................................................... 30
Results ................................................................................................................................................................... 30
Discussion ............................................................................................................................................................. 31
Conclusion ............................................................................................................................................................ 34
1 2017 APCD AQ Report
Executive Summary Ozone levels in the western portion of San Luis Obispo County remained low in 2017, though the
Atascadero, Paso Robles, and Nipomo Regional Park sites each exceeded the federal standard once this
year. The rural eastern portion of the county, which has been designated as a nonattainment zone for the
federal standard since 2012 (see Figure 1), continued to experience occasional exceedances of the
standard, with one day exceeding the standard at Carrizo Plain and six days at Red Hills. Nonetheless, a
clear downward trend in ozone levels is apparent for Red Hills, which is the site most impacted by ozone
pollution (Figures 7 and 8).
Smoke from wildfires had major impacts on ozone levels this year. Two large wildfires burning near
California–Oregon border contributed to many of the year’s highest ozone concentrations (Table 3),
including the standard exceedances at Paso Robles, Atascadero, Red Hills, and Nipomo Regional Park on
September 1st and/or 2nd. These were the Salmon August Complex Fire, which burned 65,000 acres
Siskiyou County, and the Chetco Bar Fire, which burned nearly 200,000 acres within the Rogue River–
Siskiyou National Forest in Oregon. Similarly, the July 9th exceedances at Carrizo Plain are likely related to
the Alamo Fire, which began on July 6th and burned more than 28,000 acres near Twitchell Reservoir,
and/or the Whittier Fire, which began July 8th and burned over 18,000 acres in Santa Barbara County.
South County air quality continues to be impacted by dust blown from the Oceano Dunes State Vehicular
Recreation Area (ODSVRA) along the coast. While the federal PM10 standard was not exceeded at any site
in 2017, the more stringent state standard was exceeded on more than a quarter of days on the Nipomo
Mesa, which is an increase from the previous year. In addition, the Rule 1001 performance standard was
violated 66 times. With regard to PM2.5, neither 24-hour nor the annual average standards were exceeded
anywhere in the county this year.
While windblown dust was the predominant source of high particulate matter levels in the South County,
smoke from wildfires also contributed to high particulate events throughout the county. All of the year’s
highest PM10 and PM2.5 values at Atascadero, Paso Robles, and San Luis Obispo (Table 4) were due to
either the Thomas Fire, which burned 280,000 acres in Ventura and Santa Barbara Counties in December,
or the October Northern California wildfires, which collectively burned nearly 250,000 acres. These fires
also caused some of year’s peak values at Nipomo Regional Park, Oso Flaco, and Mesa2. Construction-
related dust caused at least one exceedance of the state PM10 standard this year, specifically on December
4th at Mesa2, which was the highest recorded value for that site in 2017.
There were no exceedances of the standards for nitrogen dioxide or sulfur dioxide at any stations this
year.
This report contains two appendices. Appendix A presents a new approach for evaluating the effects of
ODSVRA dust control projects on downwind PM10. Appendix B reports the results of respirable crystalline
silica sampling conducted in 2017 and 2018 at CDF. None of the 9 samples exceeded the OSHA workplace
silica standard; however, we cannot conclude that there is no risk since our results may underestimate
actual silica levels. See the appendix for details.
Finally, there were some network outages in 2017 that affect the data in this report:
• The Oso Flaco site was temporarily shut down from December 15, 2016, through March 25, 2017.
• The PM10 monitor at the Paso Robles site (which is managed by the Air Resources Board)
experienced 3 separate outages lasting one to four weeks each. As a result, the monitor fails
annual data completeness requirements.
2 2017 APCD AQ Report
The air quality database for
San Luis Obispo County is a
public record and is
available from the District
office in various forms,
including comprehensive
records of all hourly or other
sample values acquired
anywhere in the county.
Data summaries are
published in Annual Air
Quality Reports, like this one.
Summary data appear
weekly in the Saturday
edition of The Tribune, a
local newspaper. Ambient
monitoring data is added to
separate archives
maintained by EPA and ARB.
Summary data from San
Luis Obispo County can be
found in EPA and ARB
publications and on the
world wide web at the
following websites:
www.slocleanair.org
APCD website
www.arb.ca.gov
ARB website
www.epa.gov
US EPA website
www.airnow.gov
Air Quality Index site
Air Quality Monitoring and Data
Air quality in San Luis Obispo County was measured by a network of 11
permanent ambient air monitoring stations in 2017; their locations are
depicted in Figure 1. The San Luis Obispo County Air Pollution Control District
(District) owns and operates seven permanent stations: Nipomo Regional
Park (NRP), Grover Beach, Morro Bay, Atascadero, Red Hills, Carrizo Plain,
and the CDF fire station on the Nipomo Mesa. The California Air Resources
Board (ARB) operates stations in San Luis Obispo and Paso Robles. Two
stations are owned by third parties but operated by the District: Mesa2,
located on the Nipomo Mesa and owned by the Phillips 66 refinery, and Oso
Flaco, located within the ODSVRA and owned by the California Department of
Parks and Recreation. See Table 2 for a summary of the pollutants monitored
at each station.
Air quality monitoring is subject to rigorous federal and state quality
assurance and quality control requirements, and equipment and data are
audited periodically to ensure data validity. Gaseous pollutant levels are
measured every few seconds and averaged to yield hourly values. Particulate
matter (PM2.5 and PM10) is sampled hourly using Beta Attenuation Monitors
(BAMs). All monitoring instruments are Environmental Protection Agency
(EPA)-approved Federal Equivalent Methods (FEMs) or Federal Reference
Methods (FRMs).
The 2017 data reviewed in this report was extracted from the EPA’s Air
Quality System (AQS) database. Prior to being uploaded to AQS, all data were
thoroughly reviewed and validated by the collecting agency (i.e., ARB for data
from Paso Robles and San Luis Obispo and the District for all other sites). The
raw data and the R-code used to compile the statistics and generate the
graphs in this report are available online at
https://github.com/sloapcdkt/2017aqrptR.
3 2017 APCD AQ Report
Figure 1: Map of Monitoring Stations in San Luis Obispo County
4 2017 APCD AQ Report
Table 1: Ambient Air Quality Parameters Monitored in San Luis Obispo County in 2017
O3 NO NO2 NOx SO2 PM10 PM2.5 WS WD ATM
APCD Permanent Stations
Atascadero X X X X X X X X X
Morro Bay X
X X
Nipomo Regional Park X X X X X X X X
Red Hills X X X X
Carrizo Plain X X X X
CDF X X X X
Grover Beach X X
ARB Stations
San Luis Obispo X X X X X X
Paso Robles X X X X X
Operated by APCD
Mesa2 X X X X X X
Oso Flaco X X X X
Abbreviations and Chemical Formulas:
O3 Ozone SO2 Sulfur Dioxide PM10 Particulates < 10 microns WS Wind Speed
NO Nitric Oxide CO Carbon Monoxide PM2.5 Particulates < 2.5 microns WD Wind Direction
NO2 Nitrogen Dioxide ATM Ambient Temp
NOx Oxides of Nitrogen
5 2017 APCD AQ Report
Ambient Air Pollutants Of Local Concern Ozone
Ozone (O3) is a gas that is naturally found near the earth’s surface at low concentrations, typically 10 to 40
parts per billion (ppb). It is also a principle component of photochemical smog, produced when precursor
pollutants such as volatile organic compounds and nitrogen oxides react under the influence of sunlight.
Ozone precursors are emitted by many human activities, but industrial processes and motor vehicles are
primary sources. The chemistry of atmospheric ozone is complex, and in the absence of sunlight, ozone is
destroyed by reaction with the same precursor molecules that fuel its formation during the day. As a
result, ozone concentrations typically increase as sunlight intensity increases, peaking midday or in the
afternoon and gradually declining from there, typically reaching their lowest levels in the early morning
hours and just before sunrise, as shown in Figure 2, below.
Figure 2: Example of Diurnal Ozone Pattern from Carrizo Plain
As a pollutant, ozone is a strong oxidant gas that attacks plant and animal tissues. It can cause impaired
breathing and reduced lung capacity, especially among children, athletes and persons with compromised
respiratory systems; it can also cause significant crop and forest damage. Ozone is a pollutant of particular
concern in California where geography, climate, and emissions from industrial and commercial sources
and millions of vehicles contribute to frequent violations of health-based air quality standards.
While ground level ozone is harmful to plants and animals and is considered a pollutant, upper level
(stratospheric) ozone occurs naturally and protects the earth from harmful ultra-violet energy from the
sun.
Particulate Matter
Ambient air quality standards have been established for two classes of particulate matter: PM10 (respirable
particulate matter less than 10 microns in aerodynamic diameter), and PM2.5 (fine particulate matter 2.5
microns or less in aerodynamic diameter). Both consist of many different types of particles that vary in
6 2017 APCD AQ Report
their composition and toxicity. PM2.5 tends to be a greater health risk since these particles can get lodged
deep in the lungs or enter the blood stream, causing both short and long-term damage. Sources of
particulate pollution include diesel exhaust; mineral extraction and production; combustion products from
industry and motor vehicles; smoke from open burning; paved and unpaved roads; condensation of
gaseous pollutants into liquid or solid particles; and windblown dust from soils disturbed by demolition
and construction, agricultural operations, off-road vehicle recreation, and other activities.
In addition to its harmful health effects, particulate matter can also greatly reduce visibility.
Nitrogen Dioxide, Sulfur Dioxide, and Carbon Monoxide
Nitrogen dioxide (NO2) is the brownish-colored component of smog. NO2 irritates the eyes, nose and
throat and can damage lung tissue. Sulfur dioxide (SO2) is a colorless gas with health effects similar to NO2.
Both pollutants are generated by fossil fuel combustion from mobile sources such as vehicles, ships, and
aircraft and at stationary sources such as industry facilities, homes, and businesses. SO2 is also emitted by
petroleum production and refining operations. These pollutants can create aerosols, which may fall as acid
rain causing damage to crops, forests, and lakes. They can also exacerbate asthma and harm the human
respiratory system.
Carbon monoxide (CO) is a colorless and odorless gas that can interfere with the ability of red blood cells
to transport oxygen. Exposure to CO can cause headaches, fatigue, and even death. CO results from fuel
combustion of all types, but motor vehicles are by far the chief contributor of CO in outdoor air.
7 2017 APCD AQ Report
State and National Ambient Air Quality Standards California ARB and the U.S. EPA have adopted ambient air quality standards for six common air pollutants
of primary public health concern: ozone, particulate matter (PM10 and PM2.5), nitrogen dioxide, sulfur
dioxide, carbon monoxide, and lead.1 These are called “criteria pollutants” because the standards establish
permissible airborne pollutant levels based on criteria developed after careful review of all medical and
scientific studies of the effects of each pollutant on public health and welfare.
The National Ambient Air Quality Standards (NAAQS; see Table 2) are used by EPA to designate a region as
either “attainment” or “nonattainment” for each criteria pollutant. A nonattainment designation can trigger
additional regulations aimed at reducing pollution levels and bringing the region into attainment. For most
pollutants, the NAAQS allow a standard to be exceeded a certain number of times each calendar year
without resulting in a nonattainment designation. Additionally, exceedances caused by exceptional events
(see below) may be excluded from attainment/nonattainment determinations at the discretion of the EPA.
In May 2012, the EPA designated the eastern portion of San Luis Obispo County as marginally
nonattainment for the 8-hour ozone standard. This was based on data from enhanced monitoring over
the previous decade that revealed previously unrecognized high ozone levels in that region; the western
portion of the county retained its attainment status. (See Figure 1 for the boundary between the
attainment and nonattainment areas.) In October 2015, the standard was lowered from 75 to 70 ppb, and
in April 2018, the EPA designated the eastern portion of the county as a marginal non-attainment zone for
the new standard. The county is currently designated as attaining all other NAAQS.
The California Ambient Air Quality Standards are generally more restrictive (i.e. lower) than the NAAQS,
and typically are specified as not to be exceeded. Thus, a single exceedance is a violation of the applicable
standard and triggers a nonattainment designation. As a result, San Luis Obispo County is designated as a
nonattainment area for the state one-hour and 8-hour ozone standards, as well as the state 24-hour and
annual PM10 standards. The county is designated as attaining the state annual PM2.5 standard.
State and federal standards for NO2 have never been exceeded here. The state standard for SO2 was
exceeded periodically on the Nipomo Mesa until 1993. Equipment and processes at the facilities
responsible for the emissions were upgraded as a result, and the state SO2 standard has not been
exceeded since that time. Exceedances of the federal SO2 standard had never been recorded here until
2014, when maintenance activities at these facilities resulted in emissions exceeding the 1-hour standard
of 75 ppb. (This standard was established in 2011.) State CO standards have not been exceeded in the
county since 1975. The county has never been required to conduct lead monitoring.
Exceptional Events
Exceptional events are unusual or naturally occurring events that can affect air quality but are not
reasonably controllable or preventable and are unlikely to reoccur at a particular location. Examples
include wildfires and tornadoes. Air quality monitoring data influenced by exceptional events can
sometimes be excluded from regulatory determinations related to violations of the NAAQS, if
recommended by the APCD and ARB and approved by the EPA. The APCD has not submitted any
exceptional event documentation for 2017 and does not expect any data compiled in this report to be
excluded from future attainment determinations.
1 In addition to these six pollutants, California also has standards for hydrogen sulfide, sulfate, vinyl chloride, and
visibility reducing particles.
8 2017 APCD AQ Report
Table 2: Ambient Air Quality Standards for 2017 and Attainment Status*
A standard
exceedance occurs
when a measured
pollutant
concentration
exceeds (or in some
cases, equals) the
applicable standard
prescribed by state or
federal agencies. It
does not necessarily
constitute a violation.
A standard violation
may occur following a
single or cumulative
series of standard
exceedances. Criteria
constituting a
violation are unique
for each pollutant.
A nonattainment
designation occurs
when a state or
federal agency
formally declares an
area in violation of a
standard. Typically,
ARB performs
designations
annually. Several
years often pass
between EPA
designations.
Averaging Time
California
Standard†
National
Standard†
Ozone
(O3)
8 Hours 70 ppb 70 ppb §
1 Hour 90 ppb
Respirable
Particulate
Matter
(PM10)
24 Hours 50 g/m3 150 g/m3
1 Year‡ 20 g/m3
Fine
Particulate
Matter
(PM2.5)
24 Hours 35 g/m3
1 Year‡ 12 g/m3 12 g/m3
Carbon
Monoxide
(CO)
8 Hours 9.0 ppm 9 ppm
1 Hours 20 ppm 35 ppm
Nitrogen Dioxide
(NO2)
1 Year‡ 30 ppb 53 ppb
1 Hour 180 ppb 100 ppb
Sulfur Dioxide
(SO2)
3 Hours 500 ppb
(secondary)
1 Hour 250 ppb
75 ppb
(primary)
Lead
(Pb)
3 Month 0.15 g/m3
30 Day 1.5 g/m3
* San Luis Obispo County (in whole or in part) is designated as nonattainment for the standards in boldface print as of
November 2018. † For clarity, the ozone, SO2, and NO2 standards are expressed in parts per billion (ppb), however most of these
standards were promulgated in parts per million (ppm). When comparing to the national PM10 standard, federal
regulations state that measurements shall be rounded to the nearest 10 g/m3. Thus 24-hour averages between 150
and 154 g/m3 are not considered exceedances of the standard, even though they are greater (or equal to) 150 g/m3. ‡ This standard is calculated as a weighted annual arithmetic mean.
9 2017 APCD AQ Report
Ozone and Gaseous Pollutant Summary In 2015, the federal 8-hour ozone standard was lowered from 75 to 70 parts ppb, which is the same level
as the state 8-hour standard. The old 75 ppb standard was exceeded only once in 2017: July 9th at Carrizo
Plain. Exceedances of the current federal standard occurred on 7 days countywide, with 6 days at Red
Hills2 and 1 day each at Carrizo Plain, Atascadero, Paso Robles, and Nipomo Regional Park. The state 1-
hour standard for ozone (90 ppb) was exceeded only once this year: July 9th at Carrizo Plain.
Standards for nitrogen dioxide and sulfur dioxide were not exceeded in 2017. In fact, there were only two
hours in the entire year when sulfur dioxide levels exceeded 0 ppb.
Maximum Values
Table 3 lists the highest hourly (and for ozone, 8-hour) values recorded in 2017 for ozone, sulfur dioxide,
and nitrogen dioxide at the stations where they are monitored. Concentrations are in parts per billion
(ppb). The sample date appears under each pollutant value in the format “month/day.” Values that exceed
federal standards are shown in bold,2 and those exceeding state standards are underlined.
Many of the highest 1-hour and 8-hour ozone concentrations (including standard exceedances at Paso
Robles, Atascadero, Red Hills, and Nipomo Regional Park) occurred on September 1st or 2nd. These are
likely related large wildfires that were burning near California–Oregon border, specifically the Salmon
August Complex Fire in Siskiyou County and the Chetco Bar Fire within the Rogue River–Siskiyou National
Forest in Oregon. Similarly, the July 9th 1-hour and 8-hour exceedances at Carrizo Plain are likely related to
the Alamo Fire, which began on July 6th and burned more than 28,000 acres near Twitchell Reservoir,
and/or the Whittier Fire in Santa Barbara County, which began July 8th burned over 18,000 acres.
Table 3: Highest Measurements for Gaseous Pollutants in 2017
Station O3 1-hour O3 8-hour SO2 1-hour NO2 1-hour
1st 2nd 3rd 1st 2nd 3rd 4th 1st 2nd 3rd 1st 2nd 3rd
Paso Robles 83 09/01
82 03/14
78 09/02
74 09/01
70 09/02
66 05/21
65 10/16
Atascadero 77 09/01
77 09/02
75 03/14
72 09/01
66 09/02
62 09/28
62 10/18
39 12/14
38 12/07
38 12/13
Morro Bay 71 09/02
67 10/16
66 09/03
62 09/02
60 09/03
57 10/16
52 09/26:
San Luis Obispo 74 09/02
72 09/01
67 04/05
66 09/01
66 09/02
61 10/16
59 09/26
Red Hills 79 09/02
76 09/01
76 10/17
73 2 10/17
72 09/02
72 10/16
71 09/01
Carrizo Plain 91 07/09
76 06/23
76 07/15
80 07/09
70 06/06
70 06/23
69 05/23
Nipomo Regional
Park 76
09/02
70 10/06
69 09/01
71 09/02
65 09/01
62 10/06
61 10/14
32
12/29
30 12/28
26 12/10
Mesa2, Nipomo 2
03/29
1 10/26
0 See
text
2 The 8-hour average for Red Hills from October 17th exceeds 70 ppb, but it is not an exceedance of the federal
standard under 40 CFR 50 Appendix U because it occurred in the early morning. It is still considered an exceedance of
the state standard.
10 2017 APCD AQ Report
Visual Ozone Summary
Figures 3 and 4 depict the ozone values from each station where it was monitored in 2017. The maximum
8-hour average for each day is shown for each site; exceedances of the 70-ppb standard are shown in red
with the day of month printed beside them. The heavy “stair step” line marks the monthly median. The
vertical axis extends to the annual maximum; units are ppb.
Figure 3: Daily Maximum 8-Hour Average for 2017
11 2017 APCD AQ Report
Figure 4: Daily Maximum 8-Hour Average for 2017
12 2017 APCD AQ Report
Particulate Matter Summary In 2017, there were no exceedances of the federal 24-hour PM10 standard (150 g/m3) anywhere in the
county. Exceedances of the state 24-hour PM10 standard (50 g/m3) were observed on 103 different days:
97 at CDF, 52 at Mesa2, 18 at NRP, 12 at Oso Flaco, 4 each at Paso Robles and San Luis Obispo, and 2 at
Atascadero.3 This year, CDF, Mesa2, NRP, and Oso Flaco exceeded the state annual average PM10 standard
of 20 g/m3. For PM2.5, the federal 24-hour standard (35 g/m3) and the federal and state annual average
standards (both 12 g/m3) were not exceeded anywhere in the county this year.
Local Rule 1001, which is intended to address windblown dust emissions and downwind air quality
impacts from the Oceano Dunes State Vehicular Recreation Area (ODSVRA), states that the park operator
“shall ensure that if the 24-hour average PM10 concentration at the [riding area] Monitor is more than 20%
above the 24-hour average PM10 concentration at the Control Site Monitor, the 24-hour average PM10
concentration at the [riding area] Monitor shall not exceed 55 g/m3.”4 For determining compliance with
this standard, the CDF and Oso Flaco monitors have been designated as the riding area and control site
monitors, respectively. This year there were 66 days that violated the Rule 1001 standard, as well as 10
possible violation days when the CDF 24-hour average exceeded 55 g/m3 but Oso Flaco was offline. (Oso
Flaco was not operated from December 15, 2016, to March 26, 2017.)
Note that the PM10 monitors at Paso Robles and Oso Flaco operated for only part of the year and do not
meet state and federal completeness requirements for computing annual averages.
Highest 24-hr Concentrations and Annual Averages
Table 4 lists the highest 24-hour concentrations recorded in 2017 and the dates on which they occurred, as
well as the annual means for PM10 and PM2.5 for all stations where these pollutants were monitored.
Concentrations are in g/m3. Values exceeding federal standards are shown in bold; those exceeding state
standards are underlined.
In general, standard exceedances at CDF, Mesa2, Nipomo Regional Park, and Oso Flaco are associated with
windblown dust events. This includes this year’s the top 3 PM10 and PM2.5 24-hour averages for CDF noted
in Table 4.
In addition to dust, wildfires and construction caused elevated PM10 and PM2.5 this year. The highest 24-
hour PM10 value for Mesa, recorded on December 4th, is due to construction activities in the area. The
Thomas Fire, which burned 280,000 acres in Ventura and Santa Barbara Counties in December, caused
many of the peak 24-hour PM10 values noted in Table 4: December 16th at Paso Robles, Oso Flaco, Mesa2,
and Nipomo Regional Park and December 17th at San Luis Obispo and Nipomo Regional Park. The first and
second highest PM2.5 values at Atascadero (December 12th and 28th) are also likely due to the Thomas Fire.
The October Northern California wildfires, which collectively burned nearly 250,000 acres, contributed to
the peak PM10 and PM2.5 24-hours values noted in the table for Paso Robles, Atascadero, and San Luis
Obispo. Finally, the Pier Fire, previously mentioned in the ozone discussion, contributed to elevated
3 ARB and EPA apply different conventions to the handling of significant digits. The ARB website
(http://www.arb.ca.gov/adam/topfour/topfour1.php) thus counts 82 exceedances of the state PM10 standard at CDF, 61
at Mesa2, 18 and Nipomo Regional Park, 13 at Oso Flaco, 6 at Paso Robles and 5 at San Luis Obispo. The database
used by the ARB website may also contain erroneous values. 4 San Luis Obispo County Air Pollution Control District, “RULE 1001 Coastal Dunes Dust Control Requirements,”
Adopted November 16, 2011, Revised by Court Order CV12-0013, March 7, 2016. Available online at
https://www.arb.ca.gov/drdb/slo/cur.htm.
13 2017 APCD AQ Report
particulate levels in South County on September 2nd, including the second highest 24-hour PM10 value for
Oso Flaco and the second highest 24-hour PM2.5 value for Mesa2.
Table 4: PM10 and PM2.5 Summary for 2017
Station Highest 24-hour PM10 Annual
Average PM10‡
Highest 24-hour PM2.5 Annual
Average PM2.5‡ 1st 2nd 3rd 1st 2nd 3rd
Paso Robles * 55 12/16
55 10/17
53 10/18
17.7 *
Atascadero 67 10/09
54 10/12
49 10/11 15.4 26.7
12/12 25.2 12/28
24.2 10/12
5.7
San Luis Obispo 67 12/17
63 10/09
57 10/13 16.7 25.6
10/12 23.1 10/13
21.6 10/11
6.8
CDF, Arroyo
Grande 145 04/23
138 06/10
130 03/30
38.8 32.1 06/10
30.1 05/31
29.3 03/30
9.6
Nipomo Regional
Park 101 12/16
72 06/15
68 12/17 24.9
Oso Flaco * 97 12/16
80 09/02
63 03/30
29.0 *
Mesa2, Nipomo 109 12/04
98 12/16
95 04/23 29.4 26.3
06/13 25.0 09/02
24.4 10/13 9.1
* Incomplete year, see text for details. ‡ Weighted arithmetic mean as calculated by an AMP450 AQS report.
14 2017 APCD AQ Report
Visual PM2.5 and PM10 Summaries
Figures 5 and 6, below, show the 24-hour PM2.5 and PM10 values from the stations where these pollutants
were measured in 2017. As with the ozone plots in the previous section, these show daily concentrations
by month for each site; exceedances of state and federal standards are shown in red with the day of
month printed beside them. The heavy “stair step” line marks the monthly median. The vertical axis
extends the annual maximum; units are g/m3.
Figure 5: Daily PM2.5 Values for 2017
15 2017 APCD AQ Report
Figure 6: Daily PM10 Values for 2017
16 2017 APCD AQ Report
10-Year Trends Ozone
Figure 7, below, depicts the total number of hours each year at each site during which the ozone
concentration was at or above 65 ppb. This is a useful indicator for trends, even though there are no
health standards for single-hour exposure to this level of ozone. Figure 8 shows ozone design values over
the same period. Design values are used by EPA to determine whether an area attains a federal standard.
For ozone, the design value is calculated by averaging the 4th highest annual 8-hour average over three
consecutive years. For example, a 2016 design value is the average of the 4th highest 8-hour averages from
2014, 2015, and 2016. Only design values meeting data completeness requirements are included; the
dashed red line indicates the federal 8-hour standard which changed from 75 to 70 ppb in 2015.
Figure 7: Hours At or Above 65 ppb Ozone, 2008-2017
17 2017 APCD AQ Report
Figure 8: Ozone Design Value Trends, 2008-2017
18 2017 APCD AQ Report
Particulate Matter
Figure 9 (next page) shows the number of exceedances of the state PM10 standard at each site for each
year. Collection of daily data began in mid–2009 for some sites and later for others, and years with less
than 90% valid daily data are omitted, including all years for Oso Flaco.
Figure 10 plots the total number of hours each year when PM10 was at or above 50 g/m3 during the hours
when people are most likely to be active (10 am to 4 pm). This metric is intended to illustrate trends in
population exposure, even though there are no health standards for single-hour exposure to this level of
PM10. Only years with at least 90% valid hourly data are included. Oso Flaco is not included since in 2016
and 2017 its annual data capture was only 88% and 72%, respectively; however, the annual number of
active hours at this site which were at or above 50 g/m3 were 151 and 171, respectively.
Figure 11 depicts annual average PM10 concentrations over the past 10 years;5 years with partial data are
omitted. The red dashed line marks the state standard for the annual mean (20 g/m3). While occasional
exceedances of the standard occur at most sites, the monitors on the Nipomo Mesa at Nipomo Regional
Park, Mesa2, and CDF are consistently higher than elsewhere in the county.
Trends in PM2.5 annual average are depicted in Figure 12 for the four sites where it is measured. Data for
the past 10 years are shown, and years with partial data are omitted. The red dashed line marks the 12
g/m3 state and federal PM2.5 standard for the annual mean. As with PM10, the stations on the Nipomo
Mesa tend to record higher levels than those elsewhere in the county.
Ambient Air Monitoring Network Plans The District prepares an Ambient Air Monitoring Network Plan every year. This document is an annual
examination and evaluation of the network of air pollution monitoring stations in the county. The annual
review is required by 40 CFR 58.10 and helps ensure continued consistency with the monitoring objectives
defined in federal regulations.
Each report is a directory of existing and proposed monitors in the county network and serves as a
progress report on the recommendations and issues raised in earlier network reviews. Reports also
address ongoing network design issues. They are available online at
http://www.slocleanair.org/airquality/monitoringstations.php.
As highlighted in the 2017 and 2018 reports, the Oso Flaco site was temporarily shut down by the
California Department of Parks and Recreation on December 15, 2016. The site was reopened in March
2017.
5 In general, these are seasonally weighted averages as calculated by AQS. For years when sampling methodology
changed or a site was moved, the average depicted is the time-weighted average of the methodologies or locations.
19 2017 APCD AQ Report
Figure 9: Exceedances of the California 24-hour PM10 Standard, 2010–2017
20 2017 APCD AQ Report
Figure 10: Hours At or Above 50 g/m3 PM10, 2010–2017
21 2017 APCD AQ Report
Figure 11: PM10 Annual Averages, 2008–2017
22 2017 APCD AQ Report
Figure 12: PM2.5 Annual Averages, 2008-2017
23 2017 APCD AQ Report
Appendix A: Assessing the Effectiveness of ODSVRA Mitigations Introduction
Previous year’s Annual Air Quality Reports6 contained appendices which analyzed recent trends in
particulate matter at CDF and Mesa2. They concluded that the mitigation measures deployed by the
ODSVRA operator (California Department of Parks and Recreation) in 2015 and 2016 did not detectibly
reduce PM10 levels at CDF.
This section presents new methodology for assessing mitigation effectiveness, comparing 2016 and 2017.
Since there was a relatively small change in mitigations over this period, a large change in downwind PM10
is not expected. Our real interest is in comparing 2018 and beyond to previous years, since the scale of
mitigations dramatically increased from 2017 to 2018 and is likely to continue to rise. The comparison of
2016 and 2017, presented here, is intended to “set the stage” for next year’s analysis of the ramped-up
mitigation measures deployed in 2018.
Background
Enacted by the District Board in 2011, Local Rule 10014 requires the operator of the ODSVRA to implement
dust mitigation measures with the goal of reducing PM10 emissions from the riding area of the park to the
level of emissions from non-riding areas. Prior to 2018, these efforts included temporary fencing arrays
and engineered roughness elements installed in the riding area during the windy season (April through
July), straw bale arrays in non-riding areas, and permanent revegetation in non-riding areas. All project
elements have been located upwind of the CDF monitoring station. Table A1, below, summarizes the
mitigation efforts through 2017.7
In 2018, the District and State Parks entered in a stipulated order of abatement which envisions much
more extensive mitigations than in previous years, with over 100 acres of fencing, straw bales, and
revegetation in 2018 and likely more in subsequent years.8 Rather than being focused upwind of CDF,
these mitigations are to spread throughout the ODSVRA, and thus will likely affect PM10 levels at Mesa2 as
well as at CDF.
Determining the effect of mitigation measures on downwind PM10 concentrations is difficult because while
OHV activities increase the emissivity of the Dunes, it is wind that drives the actual dust emissions. Thus,
all else being equal, windier years are expected to be dustier than less windy years. This effect can be seen
in Figures 9, 10, and 11, above. There were no mitigations projects from 2010 to 2013, yet the exceedance
count at CDF ranged from 65 to 93, likely reflecting year-to-year variation in meteorology. Similarly, the
mitigation projects for 2015 and 2016 were essentially the same, yet the exceedance counts increased at
both sites. In 2017, the size of the mitigation was cut in half, and PM10 levels increased; however, given the
previously noted inter-annual variability, it is hard to attribute this change in exceedances to the change in
mitigations.
6 San Luis Obispo County Air Pollution Control District, “2015 Annual Air Quality Report” and “2016 Annual Air Quality
Report” at https://www.slocleanair.org/library/air-quality-reports.php. 7 See District webpage, “Oceano Dunes Efforts,” at https://www.slocleanair.org/air-quality/oceano-dunes-efforts.php
for summaries of mitigation measure enacted thus far and related documents. 8 San Luis Obispo County Air Pollution Control District, “Materials Related to Petition 17-01, April 30, 2018,” at
https://www.slocleanair.org/who/board/hearing-board/actions.php.
24 2017 APCD AQ Report
Table A1: Dust Mitigations on the ODSVRA
Year Mitigation Measures
Fencing Array Straw Bale Array Other
2014 15 acres 30 acres
2015 40 acres “refresh” 2014’s array Revegetation: 6 acres
2016 40 acres Engineered roughness
element array: < 1 acre
2017 20 acres
Wind speed and direction trends for this period can be summarized using wind roses, as shown in Figures
A1–A3, which show time series of wind roses for CDF. Wind roses for April, May, and June are shown, since
these are the months when most windblown dust events occur. These events occur exclusively when
winds are out of the WNW, so it is most informative to focus on this petal of the wind roses. It is apparent
that:
• There is significant inter-annual variability. For example, Figure A2 shows that for April, the
frequency of WNW winds ranges from 20% in 2015 to 30% in 2011, and the frequency of strong
WNW winds (those with speeds greater than 12 mph) varies from about 4% in 2015 to 10% in 2011.
• In general, April and June of 2014 and 2015 were less windy than previous and subsequent years.
This trend parallels the trends in PM10 levels seen in Figures 9–11.
• May and June of 2017 saw more frequent and stronger winds from the WNW than 2016, a trend
which also parallels PM10 levels.
Thus, it appears that meteorological variability alone could explain the observed trends in PM10 levels on
the Nipomo Mesa, at least qualitatively. It is worth noting, however, that the extent of mitigations deployed
on the ODSVRA also decreased from 2016 from 2017, while at the same time PM10 levels increased. This
suggests the possibility that the effects of meteorology and mitigation on PM10 could be confounded.
Figure A1: Wind Roses for CDF in April 2010 to 2017
25 2017 APCD AQ Report
Figure A2: Wind Roses for CDF in May 2010 to 2017
Figure A3: Wind Roses for CDF in June 2010 to 2017
26 2017 APCD AQ Report
Analysis
The dust control projects from 2014 to 2017 were located upwind of CDF and appear to have been
designed specifically to reduce the PM10 levels at that monitor. These mitigation projects would not be
expected to have much impact, if any, on PM10 levels at Mesa2. Yet, as shown in Figures 9–11, the trends in
the annual average and in exceedances at CDF and Mesa2 track each other. This suggests that the dust
control projects did not affect downwind PM10 levels. It also suggests the use of a Difference-in-Differences
(DiD) approach for quantitatively assessing the effectiveness of the mitigation measures.9
Difference-in-Differences Approach
DiD is commonly used in econometrics and social sciences to assess the effects of policy changes and
interventions. In its simplest form, it compares two groups over two time periods. The first period is the
baseline, i.e. before the intervention has occurred. The second period is the post-intervention period, but
critically, the intervention affects only one group—the unaffected group comprises a natural control group.
The change in the intervention group across the periods is then compared to the change in the control
control—this is the “difference-in-differences” estimator. The statistical significance of the DiD estimator
can be determined via regression. A crucial assumption of this approach is that the intervention and
control groups are sufficiently similar and that potential unobserved confounders affect the groups
equally.
Here, we use DiD to assess the effects of mitigation on PM10 levels at CDF, using the levels at Oso Flaco as a
control. The change in mitigation from 2016 to 2017 was small—about 40 acres in 2016 to 20 acres in
2017—and its effect on PM10 is not expected to be large or statistically significant. As noted in the
introduction, the real interest is in comparing 2018 and beyond to 2016 and 2017, since the scale of
mitigations dramatically increased from 2017 to 2018. Note that Mesa2 is not a suitable control site
because it is also likely to be affected by the larger scale projects of 2018 and subsequent years. Oso Flaco,
on the other hand, should remain unaffected by the mitigation projects.
Statistical Methods10
Rather than including all 24-hour PM10 averages from 2016 and 2017, only data from wind event days were
included. This is because there are potentially several sources influencing PM10 values at these sites,
including wildfire smoke, general background PM10, aerosols transported from far away, sea spray, etc.,
but the only source that is likely to be affected by the mitigation projects is windblown dust from the
ODSVRA. On wind event days, by far the biggest contributor to PM10 levels at these sites is wind-generated
dust from the ODSVRA, and on these days high levels of PM10 are measured at both sites (with CDF
typically higher than Oso Flaco). In contrast, on non-wind event days, the other sources predominate, and
PM10 concentrations are lower and generally similar between the sites. Including data from all days dilutes
any effect of the mitigations, making it harder to detect.
For this analysis, days were considered to be wind event days if they met the criteria defined in the 2015
Annual Air Quality Report.6 In that report, a simple decision tree was developed for predicting whether an
exceedance of the state PM10 standard was expected based on meteorology. Specifically, the decision tree
used wind speed and direction data from CDF and from the S1 tower on the ODSVRA to predict PM10
exceedances. The tree was developed using data from 2011 through 2014, and it can be summarized as
follows:
9 Wikipedia, “Difference in differences,” at https://en.wikipedia.org/wiki/Difference_in_differences, accessed October
16, 2018. 10 The full details of this analysis, including a working script, are available in the GitHub repository for this report at
https://github.com/sloapcdkt/2017aqrptR.
27 2017 APCD AQ Report
24-hr average PM10 is expected to exceed 50 ug/m3 whenever:
• Wind speed at 15:00 at S1 exceeds 9.445 m/s, and
• Wind direction at 13:00 at CDF is greater than 289.5 degrees.
These rules were applied to the 2016 and 2017 dataset to generate a subset of predicted wind event days,
which was then subjected to the DiD analysis described below.
The traditional DiD analysis compares a control group to an intervention group across two time periods,
one before and the other after the intervention. In this case the observations at the sites are paired—PM10
levels at CDF and Oso Flaco are observed at the same time, and it is the difference between them on a per-
event basis that is of interest. This suggests working with the differences (or ratios) of PM10 levels between
the sites, in a manner analogous to a paired T-test. (As is well known, when observations are paired, a
paired T-test is much more powerful than an unpaired T-test.)
Since DiD analysis is a linear regression, all assumptions that usually apply to linear regression must be
met if valid inferences are to be drawn. These include the assumptions of normally distributed errors and
non-serially correlated errors. To satisfy these assumptions it was necessary to log transform the data and
to explicitly model the serial correlation as continuous first-order autoregressive process. The DiD model
that was fit was: 𝑦𝑖 = 𝛽0 + 𝛽1 ∗ 𝑦𝑒𝑎𝑟𝑖 + 𝜀𝑖
where 𝑦𝑖 = log(𝐶𝐷𝐹 𝑃𝑀10𝑖) − log (𝑂𝑠𝑜 𝐹𝑙𝑎𝑐𝑜 𝑃𝑀10𝑖)
𝑦𝑒𝑎𝑟𝑖 = 1 𝑖𝑓 𝑜𝑏𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑜𝑛 𝑖 𝑖𝑠 𝑓𝑟𝑜𝑚 2017, 0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 𝛽0 = 𝑖𝑛𝑡𝑒𝑟𝑐𝑒𝑝𝑡 ; 𝛽1 = 𝑠𝑙𝑜𝑝𝑒
𝜀𝑖 = 𝐴𝑅(1) 𝑛𝑜𝑟𝑚𝑎𝑙𝑙𝑦 𝑑𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑒𝑑 𝑒𝑟𝑟𝑜𝑟
In this model, the DiD estimator is β1, the coefficient of the year indicator.
When run on wind event days from 2016 and 2017, the analysis yielded a β1 value of + 0.136 with a non-
significant p-value of 0.363. Diagnostic plots do not reveal any issues with the model assumptions. Thus,
this analysis does not find evidence of a significant change between 2016 and 2017.
Sensitivity Analysis
That the analysis finds no difference between 2016 and 2017 is unsurprising: the mitigation projects in
2016 and 2017 were small in magnitude and didn't differ much in absolute terms: a 40 acre project in 2016
and 20 acres in 2017. Our real interest is addressing subsequent years, since in 2018 over 100 acres of
mitigation were deployed, and this number will likely increase in 2019. Thus an important question is,
"How big of a change in PM10 is needed for this type of analysis to yield a statistically significant result?"
This can be addressed through simulation. Future data were simulated by taking 2017 data and leaving the
Oso Flaco concentrations as-is, while reducing the CDF levels by 25%. The simulated data were then
compared to the actual data from 2017 using the DiD approach described above. The value for β1 value
was -0.288 as expected,11 and the associated p-value, 0.028, was statistically significant. This indicates that
11 This is the expected value for the coefficient since 𝑒−0.288 = 0.75.
28 2017 APCD AQ Report
that this methodology ought to be able identify change of at least 25% in CDF PM10 levels on event days,
after (implicitly) controlling for meteorological variability.
Conclusions
The 2017 dust control project did not appear to a have significant impact on PM10 levels downwind of the
ODSVRA. This is a qualitative conclusion, based on Figures 9–11and A1–A3. Trends in PM10 levels
approximately track in meteorology, and levels at CDF and Mesa2 continue to track each other, despite the
mitigations being upwind of CDF only.
Comparing the 2016 and 2017 dust control projects, new methodology finds no difference in their effects
on PM10 levels CDF. This is a quantitative conclusion. The projects in 2017 and earlier years were small and
temporary; the project for 2018 was much larger and permanent. We intend to use this new methodology
to compare 2018 and beyond to 2017 and see if the projects are having their intended effect of reducing
downwind PM10.
29 2017 APCD AQ Report
Appendix B: Ambient Respirable Crystalline Silica Monitoring Executive Summary & Background
Inhaling very small particles of crystalline silica is known to cause lung cancer, silicosis, chronic obstructive
pulmonary disease (COPD), and kidney disease, and may also be associated with autoimmune disorders
and other adverse health effects. To protect workers from these effects, the Occupational Safety and
Health Administration (OSHA) has set a workplace standard for respirable crystalline silica of 50 g/m3
averaged over 8-hours.12,13
In many areas, crystalline silica (in the form of quartz) is the predominant constituent of beach sand.
Respirable crystalline silica particles are at least 100 times smaller than ordinary beach sand or sand used
on playgrounds. Since the particulate matter measured downwind of the ODSVRA is derived from beach
sand, there is concern that during wind events ambient air may have high levels of respirable crystalline
silica. To address these concerns, the APCD collected 4 samples for silica analysis in 2017 and 4 more in
2018. For the sake of completeness and transparency, the results of all 8 samples will be reported here,
even though this Annual Air Quality Report is specifically for the 2017 calendar year.
None of the 8 samples collected by the APCD, or the additional sample collected by State Parks, exceeded
the OSHA standard. Furthermore, a statistical analysis of these samples suggests that the probability of a
future exceedance is negligible. While these findings are reassuring, we cannot conclude that there is no
risk. As described in further detail in the Discussion section, current silica sampling methods are not
designed for outdoor ambient sampling, and the method used in our study is known to have a downward
bias when ambient winds are high. Thus, our results may underestimate actual respirable crystalline silica
levels and the associated exposure risk.
General Considerations
• Regulatory Framework. The OSHA standard applies only to workplaces—it is not an ambient air
quality standard. Furthermore, it is enforced by OSHA; the APCD has no authority to act on
exceedances of this standard.
• Appropriateness of the OSHA Standard. The OSHA standard was developed for the workplace,
and thus incorporates assumptions that may not be adequate to protect the health of the general
population.
The California Office of Environmental Health Hazard Assessment (OEHHA) has derived a chronic
reference exposure level (REL) for respirable crystalline silica of 3 g/m3.14 A REL is a
non-enforceable health benchmark: Exposure to levels less than a REL is believed to be safe. Note
that this REL is for chronic exposure, i.e. it assumes breathing this level of respirable crystalline
silica for 24 hours each day over an entire lifetime. A single 8- or 24-hour air sample exceeding this
level is not necessarily an indication of a health risk; on the other hand, an annual average
concentration exceeding the REL may indicate a health risk.
12 Centers for Disease Control and Prevention, National Institute for Occupational Health and Safety, “CDC – Silica,
General Publications – NIOSH Workplace Safety & Health Topics.” https://www.cdc.gov/niosh/topics/silica/default.html 13 U.S. Department of Labor, Occupational Safety and Health Administration, “Safety and Health Topics / Silica,
Crystalline.” https://www.osha.gov/dsg/topics/silicacrystalline/. 14 Office of Environmental Health Hazard Assessment (2000), “Determination of Noncancer Chronic Reference
Exposure Levels. Appendix D3.” https://oehha.ca.gov/media/downloads/crnr/appendixd3final.pdf.
30 2017 APCD AQ Report
• Particle Size Fraction. Both the OSHA standard and the OEHHA REL are based on respirable
crystalline silica, which has a specific definition: roughly, the subset of crystalline silica particles less
than 4 microns in aerodynamic diameter, i.e. PM4. It is not appropriate to compare the crystalline
silica content of PM10 sample to the OSHA standard because such a sample would not be a
“respirable” sample.15 For the sake of brevity, the remainder of this appendix will use the term “silica” to
mean “respirable crystalline silica” as defined by OSHA.
Methodology
The District contracted with Forensic Analytical Services (Hayward, CA) for analysis of the silica samples.
We collected the samples on pre-weighed filter cartridges provided by the lab using their recommended
sampler, specifically the GS-3 cyclone (SKC Inc., Eighty Four, PA) operated at 2.75 L/min with a Gilian BDX II
pump (Sensidyne LP, St. Petersburg, FL). This method meets the OSHA silica rule requirements for
compliance sampling, but it may not be appropriate for sampling in high wind conditions as discussed
later.
All samples were collected at the CDF site. The sampler was attached to the roof safety railing about 3
meters off the ground and within 2 meters of the PM2.5 sampler. At the recommendation of the lab, the
target sample time was 6 hours. Samples were collected only on days when windblown dust was
forecasted to occur. In general, samples were collected from 10 am to 4 pm Pacific Standard Time, since
this is when PM10 levels tend to peak.
The analytical laboratory used NIOSH Method 7603 (Fourier Transform Infrared Spectroscopy) for
quantification of quartz, cristobalite, and tridymite, with total silica reported as the sum of these three
species. In addition, total dust mass was quantified by gravimetry. The reporting limit for total silica varied
from 8 to 10 g/m3 and for total dust from 28 to 200 g/m3.
Results
Table B1, below, presents the results of the 8 samples collected by the APCD as well as another collected
by State Parks at CDF on March 8, 2018.16 Note that State Parks’ sample used somewhat different
methodology, and in contrast to the APCD samples it was not collected during a windblown dust event.
The table provides total silica and total dust results along with their corresponding reporting limits. The
results for 3 of the 9 silica analyses were below the reporting limit; these are reported as “< X”, where X is
the reporting limit for the sample. For the total dust analyses, results for all samples except one were
below the reporting limit. The table also provides the PM10 and PM2.5 averages from CDF for the
approximately 6-hour time periods corresponding to the silica samples. Finally, the 24-hour PM10 averages
from CDF are also included.
15 International Organization for Standardization (1995), “ISO 7708:1995. Air quality — Particle size fraction definitions
for health-related sampling.” https://www.iso.org/obp/ui/#iso:std:iso:7708:ed-1:v1:en. The ISO/ACGIH/CEN convention
definition of “respirable” is actually an equation for a sigmoid shaped curved of fraction sampled versus particle size.
Particles of exactly 4 microns are sampled at 50%, with larger fractions of smaller particles sampled, and smaller
fractions of large particles sampled. Particles greater than 10 microns are essentially not sampled at all. Also see page
521of reference 14 for further discussion. 16 John W. Kelse, March 16, 2018, “DETERMINATION OF AIRBORNE CRYSTALLINE SILICA (QUARTZ) EXPOSURE AT
Oceano Dunes State Vehicular Recreation Area and CDF Air Monitoring Site, 2391 Willow Road, Arroyo Grande,
California San Luis Obispo County, California.”
http://ohv.parks.ca.gov/pages/25010/files/ODSVRA%20and%20CDF%20Airborne%20Crystalline%20Silica%20Exposure
%20Determination%20-%20March%202018.pdf
31 2017 APCD AQ Report
Table B1: Respirable Crystalline Silica Results
Date
Time
Total Silica
Concentration
(Reporting
Limit)
Total Dust
Concentration
(Reporting
Limit)
PM10
Concentration
PM2.5
Concentration
24-hr PM10
Concentration
All concentrations in g/m3
4/25/17
10:00 –
16:00
20
(10)
< 200
(200) 300 57 98
4/27/17
10:00 –
16:00
10
(10)
< 200
(200) 286 59 108
5/12/17
9:05 –
16:30
< 8
(8)
< 50
(50) 212 40 94
6/12/17
10:00 –
16:05
10
(10)
< 50
(50) 275 70 106
3/8/18 *
9:15 –
18:12
< 12
(12)
< 28
(28) 35 5 28
4/11/18
10:11 –
16:11
10
(10)
< 60
(60) 155 31 54
4/12/18
10:10 –
16:33
< 10
(10)
70
(50) 209 42 77
5/17/18
10:00 –
16:10
17
(10)
< 50
(50) 273 NA 93
5/31/18
10:00 –
16:00
10
(10)
< 60
(60) 143 34 69
* Sample collected by State Parks (see text).
Discussion
None of the nine samples exceeded the OSHA standard of 50 g/m3; however, it is possible that
exceedances occurred on days when no samples were taken. In fact, on the day that the maximum silica
concentration was observed (April 25, 2017), the 24-hour PM10 average was only 98 g/m3. PM10 levels up
to 50% higher were observed that year (Table 4), so it seems reasonable that silica levels higher than 20
g/m3 could also have occurred. Two approaches to assessing the likelihood of exceeding the OSHA
standard are described below: a univariate approach based solely on the observed silica concentrations
and correlation approach which also considers the corresponding PM10 levels.10
Univariate Approach to Assessing the Likelihood of OSHA Exceedances
A simple approach to assessing the likelihood exceeding the OSHA standard is to construct a 95% or 99%
prediction interval around the sample mean and then see if the upper bound exceeds the standard. If it
does, then this would mean there is non-trivial chance of exceeding the standard. This approach assumes
that the 8 APCD sample represent a random sample from days when windblown dust events were
32 2017 APCD AQ Report
forecasted.17 It is complicated by the fact that the data are certainly not normal, since the concentrations
cannot be negative, and by the fact that some observations are “censored,” meaning that they were
reported as simply "less than detection limit."
The silica concentrations were assumed to follow a lognormal distribution, which is a common assumption
for environmental data, and a censored lognormal distribution was fit to the data using “EnvStats”
package18 in R software environment.19 This yielded a geometric mean and standard deviation of 10.7 and
1.4 g/m3, respectively, and 95% and 99% prediction interval upper bounds of 26.6 and 41.1 g/m3,
respectively.
Neither the 95% or 99% upper bound exceeds the OSHA standard. This implies that the likelihood of
exceeding the standard on a windblown dust event data on a future forecasted event day is very low, i.e.
less than 0.5%.
Bivariate Approach to Assessing the Likelihood of OSHA Exceedances
It seems reasonable to assume that the silica and PM10 concentrations are positively correlated, i.e. that
higher silica concentrations are associated with higher PM10 concentrations. The bivariate approach takes
advantage of this and uses the PM10 data to predict silica values. In this approach, the silica concentrations
are regressed against the corresponding PM10 concentrations, and a prediction interval is constructed
around the regression line. If the upper bound of the prediction interval reaches the OSHA silica standard
at a reasonable PM10 level, this would imply that there is non-trivial chance of exceeding the standard. As
with the univariate approach, this approach is complicated by the censoring of observations and other
issues.20
Figure B1, below, is a scatter plot of the silica concentrations versus the corresponding PM10
concentrations; the 3 censored silica values—i.e. those reported as “less than reporting limit”—are shown
in red, while the others are in blue. The silica values for the censored results are not known, but they must
be between zero and their reporting limits, so for plotting purposes, a value of half the reporting limit was
used. It is not obvious from this figure whether the silica and PM10 concentrations are correlated as
expected, but this can be formally tested using non-parametric, rank-based measures of association such
as the Kendall (τ) or Spearman (ρ) rank correlation coefficients. These tests yield positive correlation
coefficients (τ = 0.36, ρ = 0.52) but non-significant p-values (0.19 and 0.15, respectively). Thus, neither test
finds evidence for a statistically significant correlation.
The lack of evidence of a correlation between silica and PM10 implies that the univariate approach is
sufficient. The bivariate approach was explored nonetheless, as it may be useful in the future if additional
silica samples are collected. The “crch” package21 in R was used to fit a censored linear regression model
17 State Parks’ sample from March 8, 2018, does not meet this assumption since it was not collected on a forecasted
event day. It was therefore excluded from these calculations. 18 Millard SP (2013). “EnvStats: An R Package for Environmental Statistics.” Springer, New York.
ISBN 978-1-4614-8455-4. 19 R version 3.4.4 (2018-03-15), "Someone to Lean On". Copyright © 2018. The R Foundation for Statistical Computing.
https://www.r-project.org. 20 In contrast to the univariate approach, this approach does not rely on the assumption that the silica samples are
randomly drawn from the population of forecasted event days; therefore, State Parks’ sample from March 8, 2018, is
included in this analysis. 21 Jakob W. Messner, Georg J. Mayr, Achim Zeileis (2016). “Heteroscedastic Censored and Truncated
Regression with crch.” The R-Journal, 8(1), 173-181.
https://journal.R-project.org/archive/2016-1/messner-mayr-zeileis.pdf
33 2017 APCD AQ Report
Figure B1: Silica versus PM10 at CDF
and generate approximate 95% prediction intervals around the fit. The slope of the regression is positive
(0.44) but non-significant (p-value: 0.07), which is consistent with the results of the Kendall and Spearman
rank correlation tests mentioned previously. Figure B2, below, depicts the linear fit, the 95% prediction
intervals, and the OSHA standard, along with the 9 data points (using the same color scheme as in Figure
B1.) For plotting purposes, the reporting limit for the silica analysis was assumed to be 10 g/m3, i.e. for
any silica concentration predicted to be between 0 and 10 g/m3 the value reported be would be “less than
10 g/m3.” Thus, the prediction intervals and linear fit are depicted as “bottoming out” at 10 g/m3. (Note
that for the 3 censored observations, a value of half the reporting limit was still used for their silica values.)
The main feature of Figure B2 is that the 95% prediction interval does not include the level of the OSHA
standard until PM10 levels are over 930 g/m3. In other words, the model predicts that PM10 levels would
need to be greater than 930 g/m3 before there is even a small chance (2.5%) of exceeding the OSHA silica
standard. The district has never observed hourly PM10 levels this high at CDF, so it is extremely unlikely
that there have been unobserved exceedances of the OSHA silica standard at this site.
Given the degree of extrapolation in Figure B2, the exact location of the intersection of the OSHA standard
and the prediction interval should be taken with a grain of salt. A small change in the estimated slope or
residual standard error could shift the location substantially. Furthermore, this approach assumes a linear
relationship between PM10 and silica (and with only 9 data points, it would be hard to justify exploring
more complicated models), but the true relationship may be non-linear. Finally, we have pooled our
samples with State Parks’, despite differences in methods. This is another potential source of error.
(Omitting State Parks’ sample from the analysis yields a lower value (785 g/m3) for the intersection of the
prediction interval and the OSHA standard, but the overall conclusion remains the same.)
34 2017 APCD AQ Report
Figure B2: Calculation of Probability of OSHA Exceedance
Sampling Considerations
The OSHA silica rule applies to occupational settings which tend to be indoors or underground and where
wind speeds are low; thus, samplers meant to comply with the rule are usually only tested under no- or
low-wind conditions (up to 1 m/s). We used the GS-3 cyclone for this study, which is one of the few
samplers whose performance has been characterized under higher wind speed. Its efficiency was found to
decrease as winds increased; at 9.1 mph, which was the maximum wind speed tested, the mean bias
was -9.55% (range: -3 to -20%).22 To the best of our knowledge, this is the highest wind speed any silica
sampler has been tested at.
When our silica samples were collected, wind speeds were much higher, with hourly values typically
ranging from 10 to 15 mph and with gusts even higher. For this reason, it is likely that our silica samples
under estimate the actual levels. (In contrast, the EPA-approved samplers used by the District for PM10 and
PM2.5 have been extensively tested under high ambient wind speeds, and there are no concerns about bias
with these methods.)
There is some empirical evidence for under-sampling in our silica dataset. The total dust result for the
June 12, 2017, sample was less than the reporting limit of 50 g/m3, yet the PM2.5 concentration for the
same 6-hour period was 70 g/m3. The “respirable” size fraction is PM4, so by definition the total dust
concentration should be larger than the corresponding PM2.5 concentration (and smaller than the PM10
concentration), but in this sample it is significantly less than the PM2.5 concentration.
Conclusion
There were no exceedances of the OSHA silica standard among the 9 samples collected at CDF by the
APCD and State Parks. Furthermore, statistical analysis of these samples suggests that the probability of
an exceedance on a non-sampled day is negligible. While these findings are reassuring, we cannot
22 Mridul Gautum and Avula Sreenath (1997). “Performance of a respirable multi-inlet cyclone sampler.” J. Aerosol. Sci.,
28(7), 1265-1281.
35 2017 APCD AQ Report
conclude that a risk does not exist since these samples may underestimate actual respirable crystalline
silica levels. Furthermore, it may not be appropriate to compare ambient silica concentrations to the OSHA
standard since it was not intended to apply to ambient concentrations.
The District and State Parks are considering collecting additional silica samples in 2019 using different
methodology. Ideally, we would use a “respirable” size fraction sampler that has been shown to perform
well in high ambient winds; however, no such samplers are commercially available to our knowledge.
Instead, pending the availability of equipment, we hope to use sequential FRM samplers to collect PM10
and PM2.5 samples on Teflon filters. A subset of these will be quantified for crystalline silica. Such samples
will not be directly comparable to the OSHA standard, but they should be unbiased since the sequential
FRM sampler has been designed and proven to perform well even under high ambient winds.